Analyze This. Classify That.
Jules: You know what they call a Quarter Pounder with cheese in France?
Jules: Tell ‘em, Vincent.
Vincent: A Royale with cheese.
Jules: A Royale with cheese! You know why they call it that?
Brett: Because of the metric system?
Jules: Check out the big brain on Brett!
I confess. Pulp Fiction continues to be one my favorite flicks till date. The reason I quote the above conversation from it is to drive home a different point. When on the wrong side of a loaded gun (as in the scene), chances are you might just get the right answers for any purchase queries.
It’s a different story on an ordinary day in the enterprise. Consider an organization with purchasing teams in 25 countries across four continents, buying through 15 different source systems from 10,000 plus suppliers. Chances in this case are that the purchasing department couldn’t tell if they have been purchasing the same item under different names. Goodbye spend analysis.
The starting point for a spend analysis program is to gain visibility into goods and services being procured within the organization. And that seems like the easy part. After all, it’s only a matter of extracting the data from ERP, right? Wrong. The hard fact is that purchasing information often resides on disconnected islands.
And even if with some luck all the data is aggregated, the pain of extracting meaningful insights awaits. That’s because purchase data in its raw form is highly unstructured and riddled with varying taxonomies. And adding on to this is the lack of standardized classification practices which makes treading in to spend analysis territory a nightmare.
It is this context that classification and enrichment of spend data assumes a pivotal role. The ability to accurately classify purchase data into categories helps uncover significant value that lies untapped in purchasing organizations. For starters, it answers fundamental questions on buying patterns in the organization such as categories of goods and services being procured, spending across multiple categories, and overlaps in spending.
Classifying spend, however, is easier said than done. Disparate data systems and limited data enrichment capabilities inevitably play the spoilsport in every well-intended spend analysis initiative. Furthermore, the enterprise predilection for manual effort intensive processes does not help the cause much either.
The implications of such approaches are far reaching. Be it missed purchase optimization opportunities, bulk order discount leakages or inventory overheads, organizations fail to see what’s staring on their faces, let alone the hidden inefficiencies.
Getting to the holy grail of good data requires the coming together of technology, expertise and processes. A good spend classification and enrichment methodology needs to bring together the best of many worlds. Like blending automated classification techniques such as rationalization, machine-based learning, and inference models with manual audits and process checks for relevancy and accuracy. Or laying out tailored taxonomies based on customer preferences on top of industry standard frameworks for in-depth hierarchical visibility on goods and services purchased.
While powerful analytics and user-friendly interfaces are perquisites for good spend visibility, overlooking the tenets of good data is to put the cart before the horse. After all, it’s garbage in garbage out.